“DON’T PANIC, How to End Poverty” is a one-hour documentary film produced by Wingspan Productions for This World on BBC2 . Presently BBC makes the film available in UK only. Outside UK, there are two trailers on Youtube. The presenter, Professor Hans Rosling, is co-founder of the Gapminder Foundation which supplied the data and the underlying visualizations used in the film.
Below are the data sources used for the visualizations in the film, sorted by order of appearance. Where possible, the links to data sources are given as well as a brief summary of Gapminder’s work of merging, curating and rounding numbers for visual clarity.
Explore the data bubble chart with Child mortality and GDP per capita in this Interactive Bubble Chart.
Explore the income distributions of all countries in this interactive tool.
General note
At Gapminder we are continuously updating our data. Hence, some of the data you see might be based on an earlier revision. We compile and curate data to make major global trends easy to understand. We strongly advise against using our compiled and curated data from different sources for numeric analysis or official purposes. Reason being that we do considerable simplifications and gap-filling in order to facilitate understanding, We welcome critique, comments and advice about any data we have used or not used. We invite people to suggest improvements and to contribute with additional sources.
We think of our work as similar to the early works of cartographers. The first world maps were considerably wrong when it came to many details, but they still provided a big picture which hadn’t existed before. Over time quality global statistics improve thanks to careful revision, just like the world maps got better and better. Think is true also for income distribution data. The research on world income distribution is still in an early stage, and we still don’t have representative data for all countries. However the quality of the picture is gradually improving thanks to with higher granularity of this very important data. But a coherent big picture of the changing pattern of world income distribution has already emerged which is consistently reproduced in a wide range of alternative sources.
1. Mapping extreme poverty on the spinning globe
The most recent estimate of the total the number in extreme poverty is: 836 million people. We show it as 84 dots on the world map, each representing 10 million people. The number of dots per country was decided by rounding the numbers from World Bank’s Global Monitoring Report 2014/2015, table 1, page 19. The positions of dots within countries were guided by Poverty Maps from two sources: World Bank & Worldpop. The underlaying map was prepared by Max Roser and Hans Rosling.
2. Income Distribution in Sweden
The yellow bell curve shows the position of the Swedish population on The Adjusted Global Income Scale described below. The relative poverty line of Sweden is positioned based on the official definition of “Relative Low Economic Standard” in Sweden “A household with a disposable income, adjusted with consideration to the dependency burden, lower than 60 percent of the median value in the population.“ The median income is horizontal position where you find the top of the bell curve. We draw the Swedish line at 60% of the median income. Here are two reports from Statistics Sweden discussing the relative poverty in Sweden: Increasing relative poverty rate(2012) The increasing number of people at risk of falling below the line (2013)
Sweden’s horizontal position on the total scale is determined by the World Bank estimate for GDP per capita in PPP 2011, from World Development Indicators: 44029 $/year for the year 2014. We extrapolated the number to year 2015 using the IMF projections in World Economic Outlook April 2015 edition. The width of Swedish bell curve shows the distribution of people on different income levels, based on the most recent Gini available from Eurostat SILC: 24.9 for year 2014, “Gini coefficient of equivalised disposable income.” We use this Gini from 2014 for the year 2015 as the change in Gini for Sweden is expected to be extremely small.
3. Income Distribution in Malawi
The blue bell curve shows the position of the Malawi population on the Adjusted Global Income Scale described below. The extreme poverty line on this scale is 1.85$/day as described below. Malawi’s horizontal position is determined by the World Bank estimate for GDP per capita in PPP 2011, from World Development Indicators: 784 PPP$/year, for the year 2014. We extrapolated the number to year 2015 using the IMF projections in World Economic Outlook April 2015 edition. The width of Malawi’s bell curve shows the distribution of people on different income levels, based on the most recent Gini avaialble: 45.2 for year 2011. This Gini comes from the dataset The UNU-WIDER World Income Inequality Database (WIID) (v3.0B). It is the most recent number prefered by Milanovic in “All the Ginis”, defining it as representative for the whole population, measuring the distribution of households based on their Per capita consumption. We use this Gini from 2011 to show Malawi in 2015, as the distribution of income has most likely not changed much.
4. The Extreme Poverty Line
The line for extreme poverty is set to 1.85$ per day in PPP 2011. It differs from the 1.25$ /day which is in PPP2005. The recent official Poverty Line of the World Bank and UN is 1.25$/day adjusted for international prices in year 2005. But prices change. Our slightly higher level for the poverty line is an adjustment for the 2011 price comparisons like the rest of the graph. The World Bank recently published a new round of global price comparisons from the International Comparison Program; called PPP 2011; meaning Purchasing Power Parity in prices of year 2011. We use those for our GDP per capita series and income scales throughout the film. We expect the UN & the World Bank will soon update the official Poverty Line to these recent price levels based on this Working paper. Meanwhile we have positioned the new poverty line at 1.85$ per day according to the new PPP 2011.
5. Dollar Street
The photos and videos come from the Dollar Street project which is a visual teaching framework for understanding how everybody lives. It was invented by Anna Rosling Rönnlund who is also the Product Manager of Gapminder. The project will launch later in 2015. To learn more watch this TEDx talk about Dollar Street. To not miss the launch you can follow Dollar Street on facebook.
6. Global Poverty Rate Long Trend
The long trend for global poverty rate is a combination of three data sources:
— Source for 2015: UN official Poverty rate
The official MDG website express the official UN estimate vaguely as “At the global level more than 800 million people are still living in extreme poverty.” here. The precise estimate is available on page 15 in the MDG report for 2015: 836 million people below 1.25$/day globally. We divide this estimate with the UN official estimate of the present world population: 7 349 million people. 836/7349 = 11.3% of world population are in extreme poverty.
— Source for 1981 to 2011. World Bank, PovCalNet
Estimates of global Poverty Rate are available from World Bank’s PovCalNet, or here. We use the indicator “Poverty headcount ratio at $1.25 a day (PPP) (& of population) , which is defined as “Population below $1.25 a day is the percentage of the population living on less than $1.25 a day at 2005 international prices.“
— Sources before 1981: Economic History Researchers
We aligned our trend with two historical sources that have very similar estimates.
— “Inequality Among World Citizens” 1820–1992 By Francois Bourguignon and christian morrisson FRANCOIS BOURGUIGNON AND CHRISTIAN MORRISSON
— Page 40 in “The changing shape of Global Inequality – exploring a new dataset” (Zanden, Baten, Foldvari, Leeuwen)
We have recreated the poverty rates of these historical sources, by using the same method, but relying on more recent estimates of historic GDP per capita and added estimates for missing countries and years. This is further described in the section Adjusted Global Incomes below. This work was done jointly by Hans Rosling, Ola Rosling & Max Roser. You can see the alternative series here: OurWorldInData
7. Global Population Trend
Data for the period 1950 to 2015 comes from UN Population Division. Data for the period 1950 to 2015 comes from UN Population Division. Estimates before 1950 comes primarily from International Historical Statistics compiled by Mitchell and the Maddison Project, with many minor additions by Gapminder. We sum up the population estimates of all countries to get the total World population.
8. World Income Distribution
You can explore the income distributions of all countries in this interactive tool.
This graph is constructed by combining data from multiple sources. In summary, we take the best available country estimates for the three indicators: GDP per capita, Population and Gini (which is a measure of income inequality). With these numbers we can approximate the number of people on different income levels in every country. We then combine all these approximations into a global pile using the method described below under The Adjusted Global Income Scale.
9. World Quiz — Correct Answers
The sources for the data behind each of the three questions are given below:
— Question 1 “How many people out of 10 have electricity at home? The indicator used is “The percentage of population with access to electricity”. The data is collected by the World Bank from industry, national surveys and international sources. The latest estimate for the World is 83% for the year 2010. Historical data for access to electricity are estimated based on this documentation from the International Institute for Applied Systems Analysis (IIASA) in Austria.
— Question 2 “How many children out of ten are vaccinated against measles? The indicator is “The percent of children that received one dose of measles vaccine by their first birthday”. One dose is enough to protect against measles if the child gets it before their first birthday. Data is compiled by the World Health Organization (WHO).
— Question 3 “How many girls out of 10 go to primary school? This indicator is called “Adjusted net enrollment rate, Primary” and measures the percent of primary school-age girls that attend primary school. The age-range and length of primary school varies between countries. The data is provided by the World Bank and the latest estimate for the world is 90% in 2013. Historical data for the percent of girls in the World that attended primary school are roughly extrapolated backwards using the historic trends in literacy summarized here “Our World in Data”.
10. Child Mortality vs. GDP Per Capita
You can explore the data in this bubble chart Interactive Bubble Chart.
The vertical axis shows an indicator called Child Mortality Rate: “Yearly mortality under age five, per 1000 live born”. In this film it is expressed as %. The primary data source is The UN Inter-agency Group for Child Mortality Estimation. Their trend data cover most countries back to 1950 and for several countries all the way back to 1930. The historical stats before 1950 are compiled by Mattias Lindgren at Gapminder, from many different historical sources, documented here. We extrapolated the series up to 2015 using the projections from UN Population Division.
The horizontal axis shows GDP per capita in Constant International Dollars which are adjusted for price differences between countries and price changes over time. The main data source is the World Bank’s GDP per capita in PPP 2011, from World Development Indicators, which is based on PWT 8.1. Those trends often end in 2013, and we extrapolated them up to year 2015 using the IMF projections of World Economic Outlook April 2015 edition. The World Bank series only goes back to 1980. Before 1980 we have linked the historic growth rates from the Maddison Project, as well as additional trends from Gapminder’s historic GDP per capita estimates.
The size of the Bubbles show the total population in each country. Data for the period 1950 to 2015 comes from UN Population Division. Estimates before 1950 comes primarily from International Historical Statistics compiled by Mitchell and the Maddison Project, with many minor additions by Gapminder.
11. Aid Levels
The aid levels in the film are based on estimates of the amount of international development assistance per person in extreme poverty in different country income groups. The data-source is “Financing the Future” from Overseas Development Institute, ODI” and the report as well as the datasets are available online.
12. Family Size by Income
People below the extreme poverty do not live only in the poorest countries. Some live in middle income countries. To estimate the combined Fertility Rate of the poorest people, we can not just use country averages. We used data that divide countries into five different groups based on their economic resources; so called wealth quintiles. The underlying data is available here: WHO Health Observatory, which derived those estimates from DHS surveys. We assume that the people in each wealth-quintile coincide with the people in the income quintiles from Income Surveys provided by World Development Indicators: such as 20% poorest part of total GDP. Those were used to express the Income per Capita of the people in each wealth-quintile and work was done by Max Roser, who published the quintile plot here. For some countries this kind of granular data is missing and we couldn’t extract the poorest part of the population to our sum. We believe the estimate 5 children per family, is actually lower than what we would have gotten if quintile data had been available for more countries.
The Adjusted Global Income Scale
Background
When comparing people’s economic standard of living within a country, it’s preferable to compare disposable income after social transfers (such as taxes and welfare programs). But for international comparisons, those national-measures are not simple to use because welfare-systems and tax-systems differ widely across countries. So far, there exist no single data source that adjusts for the wide differences across a majority of countries. Instead, when investigating long trends of global income distribution, researchers are using less accurate measures which are more comparable. The only economic indicator which is widely available and sufficiently standardized to allow for historic international comparisons is GDP per capita.
Preferably the global income distribution should be visualized based on comparable global household surveys with a comparable measure for disposable income, adjusted not only for differences in cost of living between countries, but also for differences within countries. Unfortunately no such income survey with global coverage exists, not even for a single year. We strongly agree with Branko Milanovic about the urgent need for such comparable surveys, especially for the purpose of monitoring the UN poverty-goal. In the absence of such survey, we and others are forced to rely on less reliable methods.
But it’s still a reasonable alternative. Here’s a comparison by Martin Ravallion. He concludes that the GDP-per-capita-based-method, which we use, actually results in a very similar general poverty trends as the more accurate survey-based-method. The different methods yield income results that are different in terms of income level, but they are similar in terms of the percentage change over time.
The GDP-based method summarized
The Global Income Distribution graph is basically an animating clone of static graphs from the leading researchers in this field of Economic History: “The changing shape of Global Inequality – exploring a new dataset” (Zanden, Baten, Foldvari, Leeuwen). We use the same method, but with more up-to-date numbers. The method goes like this:
— Step 1. We use three data points for each country and year: Population, GDP per capita and Gini (Gini expresses how skewed the income distribution is within a population; See Wikipedia).
— Step 2. We assume the distribution is log-normal in every country and year, which means the population pile up like a bell curve on a logarithmic income scale. This assumption is common among economist using this method, and it’s surprisingly solid when compared to empirical data.
— Step 3. We draw the bell curve for every country on a global logarithmic income scale. The width of the curve depends on the Gini and horizontal position depends on the GDP per capita. The areas of the bell curves are relative to the population of the country.
— Step 4.We stack the shapes of the country bell curves on top of each other for each region, and then we stack the regions on top of each other and we get one a single global shape showing everybody in the world by income level.
Gapminder’s Level-Adjustments
Zanden’s group has kindly shared their method and data with us. We have then extended their data series beyond 2000, up to year 2015 using primarily Ginis recommended by Milanovic in “All the Ginis” . Zandens group uses a GDP per capita in PPP 2005. We switched to use GDP per capitas in PPP 2011 from World Development Indicators, which are based on the PWT 8.1. We then adjust the old GDP per capita levels by applying the growth rates from the Maddison Project so they are lifted to the new PPP levels of 2011. The World Bank series mostly ends in 2013 or 2014 and we have extrapolated them up to year 2015 using the IMF projections in World Economic Outlook April 2015 edition. To reach up to 2015 we have used the latest available gini from all countries and this gives us an “updated version” of the graph in Zanden’s paper.
As Martin Ravallion discusses the GDP-based numbers are comparable in terms of trends to the survey-based numbers. But the levels are systematically higher. In this presentation we wanted to give a coherent story that required combining data from the two different methods;
— 1. official trends of poverty rates originating from income surveys.
—2. global income distributions based on GDP per capita.
Therefor we adjust the GDP-based data to be aligned with survey-based sources which are more accurate. This adjustment are done in two steps:
— 1. Anchoring below the poverty line: We adjust all income levels so the people below the poverty line are aligned with the survey based method. The global income scale from the GDP-based method, is moved so everybody’s incomes get lower and 11.3% of world population are below the extreme poverty line of 1.85$/day (in PPP 2011) in year 2015.
— 2. Anchoring above the poverty line: We stretch the income-scale so the distribution of the richer people are better aligned with the survey-based method. We did this by visually comparing to the most accurate up-to-date survey-based picture of world income distribution, namely the graph for 2008 in figure A.2 in “Global Income Distribution: From the Fall of the Berlin Wall to the Great Recession” ( Lakner & Milanovic, 2015).
Interact with the Income Mountain Tool to see how different countries relate to each other. But please beware the warning sign about DATA DOUBTS and stay tuned to our continuous data improvements!
Stockholm 23 September 2015 by Hans Rosling & Ola Rosling.
Questions: [email protected]
Feedback: mailto:[email protected]
Blue countries are first or second place in the group (in the qualifying position) and red means no qualification halfway into the qualifying leg. 
On 24 April 2010, professor Angus Maddison passed away at the age of 83. Maddison was an economist and economic historian and a pioneer in exploring the broad developments of the world through statistics. 
Click here to download Gapminder Desktop
Hans Rosling, founder and director of Gapminder, is one of the world’s “100 most important global thinkers” of 2009, according to Foreign Policy Magazine.
At TEDIndia, the first ever TED-conference in Asia, Hans Rosling spoke about the rise of Asia and showed how the region has developed over the last 150 years.
Which country has the largest grain production? Who produces the most fruits, vegetables, nuts, coffee beans or other crops? And who brings up the most sheeps, cows, or other livestock?